Paper to be presented at The Third Intentional Conference ... · Evidence from Three Parliamentary...
Transcript of Paper to be presented at The Third Intentional Conference ... · Evidence from Three Parliamentary...
Party Government and Policy Responsiveness
Evidence from Three Parliamentary Democracies
Paper to be presented at
The Third Intentional Conference on Public Policy (ICPP),
Singapore, 27-30 June 2017
Lars Mäder
University of Copenhagen
Anne Rasmussen
University of Copenhagen &
Leiden University
Dimiter Toshkov
Leiden University
2
Abstract
Does party government moderate the responsiveness of public policy to public opinion? Analyzing
a new dataset we examine whether the ability of governments to respond to the public on 306
specific policy issues in Denmark, Germany, and the United Kingdom is affected by the extent of
coalition conflict and on the fit of the considered policy changes with government preferences. We
find a systematic but relatively weak positive impact of public support on the likelihood and speed
of policy change. Contrary to expectations, higher coalition conflict is not associated with fewer
policy changes nor with weaker responsiveness to public opinion. Instead, changes that would move
policy in line with the general policy preferences of the government are marginally more likely to
be adopted. But we find no evidence that responsiveness to public opinion is weaker for policy
changes that go against the preferences of the government.
3
Introduction
In democracies public policy should reflect the wishes of the people. In representative democracies
embodied in parliamentary political systems, however, public policy is made not directly by the
people, but by elected representatives in legislative assemblies and cabinet governments. In general,
regular, free and fair elections ensure that legislators and cabinet ministers have incentives to follow
the policy wishes of the voters. But legislators and ministers are not passive transmitters of public
opinion. First, they are typically members of political parties, and political parties need to govern
responsibly (Bardi et al. 2014) and cater to the interests of their members, sponsors, and voters;
thus, not necessarily to the positions of the median citizen. Second, in parliamentary systems parties
often govern in coalitions. Coalitions might make it more likely that a broad set of public interests
are represented in government, but they also erect significant hurdles to the adoption of new
policies (Tsebelis 2002) and dilute perceptions of political responsibility (Duch et al. 2015). Third,
voters do not hold strong, well-formed opinions on all policy issues, and they do not follow closely
policy developments on all issues all the time (Achen and Bartels 2016). This provides leeway to
legislators and ministers to choose when and on which issues to respond to the public. Altogether,
the patterns of party government in parliamentary democracies have considerable potential to
moderate the representation of public preferences into public policies1.
There are large literatures scattered across several subfields of political science and sociology
that have studied policy representation, on the one hand, and the influence of party government on
public policy, on the other. Research on policy representation (for recent reviews, see Wlezien and
Soroka 2016; Erikson 2015; Shapiro 2011) has established that both in the US and in Europe public
1 Political parties and elites also shape and lead rather than only follow and mediate public opinion (Achen and Bartels
2016).
4
opinion has a strong but far from deterministic influence on public policy (see also Monroe 1998;
Page and Shapiro 1983). This influence is manifested in relatively high degrees of static congruence
(for a review of the concept, see Wlezien 2016) between what the median citizen wants and what
the state of policy is (Lax and Phillips 2012), as well as in dynamic responsiveness, in which the
state of public policy adapts to shifts in public preferences (Bevan and Jennings 2014; Jennings and
John 2009; Wlezien 1995; Erikson et al. 2002). Researchers have also looked into the institutional
factors that promote or inhibit responsiveness (Wlezien and Soroka 2012; Hobolt and Klemmensen
2008), but with inclusive results: there does not seem to be a single institutional recipe that ensures
a high degree of correspondence between what citizens want and government delivers .
The literature on party government has examined the impact of party preferences on aggregate
measures of the policy output of governments. Surprisingly, there is only scant and contested
evidence for the effects of party positions on the direction of public policy shifts and the content of
policy changes (Imbeau et al. 2001). Researchers have also investigated how the scope of
differences in preferences between parties in government coalitions (Strøm et al. 2010) and between
different institutional actors within the political system (Martin and Vanberg 2011; Tsebelis 2002)
affect legislative production and policy output, and have found mostly negative effects (Saeki 2009;
Schermann and Ennser-Jedenastik 2014; Bräuninger et al. 2015; Tsebelis 1999). Yet, very few
studies of the policy effects of government preferences, coalition conflict, and related concepts,
such as veto players and preference heterogeneity, take into account the possible influence of public
opinion (but see Toshkov 2011).
It is theoretically plausible that public opinion and party government are not only competing
drivers of policy change, but that they are also entangled in a complex relationship in which parties,
legislators, and ministers shape, mediate, and moderate the influence of public opinion. Regrettably,
scholars have rarely examined the effects of public opinion and party government on public policy
5
simultaneously, especially when it comes to the case of parliamentary democracies. The existing
empirical studies of the US suggest that public opinion, political preferences, and institutional rules
interact in affecting the policy productivity of governments (Binder 1999; Coleman 1999), but these
insights have not been translated and systematically tested in the context of parliamentary systems.
In this article we focus on the question how patterns of party government moderate policy
responsiveness in three European parliamentary democracies: Denmark, Germany, and the United
Kingdom (UK). We employ a comparative design in which we track a total of 306 policy issues
across the three countries. For each of these issues, we identify measures of public opinion from
representative national opinion polls, and we trace whether policy change on the issue occurred
within a four-year period starting at the time of the opinion poll. We construct policy-specific
measures of the preferences of the governing parties, and derive from them estimates of the conflict
between coalition cabinet partners. We model two central aspects of policy making – the occurrence
of policy change and the time it takes for change to occur. The timing of policy change has not
played a prominent role in existing opinion-policy research, even though responsiveness is not only
about whether politicians adopt policies in line with what the people want, but also about how fast
they enact policy changes.
Analysing this data, we find that in all three countries public support significantly increases
the likelihood of policy change and significantly decreases the time until change is enacted. Yet,
whereas these effects are systematic, they are relatively small in substantive terms. Despite the
possible danger that governments interested in sticking to their parties' positions or experiencing a
high degree of coalition conflict may have higher difficulty responding to public opinion, we do not
find evidence that these two factors weaken the opinion-policy linkage. Contrary to our theoretical
expectations, we find that coalition conflict does not have a direct effect on the likelihood of policy
change (nor on the time until it occurs), and that there is no moderating effect of coalition conflict
6
on policy responsiveness as such. Furthermore, even though we encounter some relatively weak
support for a positive effect of government support for a policy change on the likelihood of the
change being adopted, there is no evidence that lack of support for a policy by a government
decreases responsiveness to public opinion. From a normative point of view our findings can be
interpreted as casting less of a negative view on the role of partisan preferences and coalition
conflict on responsiveness than expected. However, they also underline that the complex
relationship between patterns of party government and responsiveness may not be captured well by
existing theory.
Public opinion, party government, and policy making in parliamentary democracies
A high degree of correspondence between the policy-relevant preferences, opinions, and attitudes of
citizens, on the one hand, and the state of public policy, on the other, is a central, if not the central,
normative requirement of liberal and democratic government (Przeworski 2010). The normative
ideal of policy responsiveness is not absolute, and good reasons can be given why ‘pandering’ to
public opinion might not be desirable in any single, particular case (e.g. Weissberg 2002).
Nevertheless, no system in which there is systematic incongruity between what the public wants
and what the government delivers can be defined as democratic.
Policy responsiveness is relevant in normative terms for direct and representative democracies
alike, although the mechanisms through which it is organized in the latter are different. In
representative democracies the influence of public opinion is typically exercised through a process
of delegation in which the citizens elect representatives who share, express, and defend their views,
and then the political representatives enact policies in line with the views that enjoy the support of
the citizens. In the particular case of parliamentary democracy, political parties take central place in
the process of representation. In addition to the electoral connection, mechanisms of political
7
accountability ensure that political representatives, whether party members or not, have the
incentives to reflect the views of the citizens in the process of government (Öhberg and Naurin
2016; Müller 2000). To be sure, representation is not a passive process: representatives help
construct, articulate, and change the policy preferences of the represented, and sometimes need to
go against the explicit wishes of a majority of the public for the sake of the common good and
responsible governance (Bardi et al. 2014). But, again, gross and sustained mismatches between the
wishes of the citizens and the policy actions of their representatives would signal a failure of
representative democracy.
Empirically, there are similarly good reasons why we should expect to observe a link, albeit
an imperfect one, between public opinion and public policy. First, in the electoral contest for
political power, parties and individual candidates who express policy views more popular with the
voters have a higher chance of being elected and hold office. In theory, in electoral systems based
on majoritarian rules (plurality, single-member electoral districts), we should expect the winning
party to represent the median citizen, at least on the policy issues of electoral significance2. In
proportional electoral systems, the government as a whole, even if composed of different parties
which represent particular groups within society, is also likely to represent a wide range of interests
and preferences. Often, the median or compromise position of the governing parties would
approach the median position of the citizens (Powell 2006; Coman 2015; McDonald et al. 2004;
Blais and Bodet 2006) making policy responsiveness likely. Once elected, legislators and ministers
are not bound by a narrow mandate to represent literally the policy preferences of their voters. But
2 In practice, issues such as limited party offerings and distortions of vote-to-seat transformations hamper the process of
representation in majoritarian, single-member districts electoral systems (McDonald et al. 2004; Powell and Vanberg
2000).
8
they still have incentives not to stray too far away from the opinions of the electorate. Otherwise,
they risk damaging the party brand and losing the confidence of their voters.
Responsiveness to public opinion may be less likely in a parliamentary than a presidential
system (see e.g. Hobolt and Klemmensen 2008; Soroka and Wlezien 2010). But the risk of electoral
loss should still give elected politicians incentives to pay attention to public wishes. Moreover, the
risk that the public expresses its discontent through protests or civil disobedience should also
encourage politicians to be attentive to the views of the public (Brooks and Manza 2006), in
parliamentary and presidential systems alike.
It may be a fact that people do not hold strong, well-formed opinions on all policy issues and
do not pay close attention to policy developments all the time (Achen and Bartels 2016). However,
even if people do not have stable and meaningful absolute preferences about a policy, they might
still have reasonably clear and politically consequential relative preferences about the direction and
scale of desired policy change (Wlezien 1995). Hence, expressed public support for a particular
policy alternative sends a strong signal and can only be ignored at a significant political cost.
Therefore, our first set of theoretical expectations states:
H1: The stronger the public support for a policy alternative, (a) the higher the likelihood of policy
change and (b) the faster policy change will occur.
In parliamentary democracies public opinion is only one force of policy change amongst others, and
its influence is mediated and moderated by the political system. In the discussion below we theorize
about the expected direct and moderating effects of a set of variables that we collectively refer to as
‘patterns of party government’. These variables have to do with (1) the extent of coalition conflict
9
within cabinets, which in turn is related to the type of government (single-party or coalition;
majority or minority) and (2) the general policy preferences (positions) of the government.
In coalition cabinets, which are a typical outcome of proportional electoral systems but occur
under majoritarian rules as well, several different parties participate in government. It has been
argued that coalition governments face significant hurdles in adopting new policies and legislation,
because these have to win the consent of the coalition partners (see e.g. Martin and Vanberg 2011;
Strøm et al. 2010). Theoretically, the set of policies that is preferred to the status quo decreases with
increasing preference heterogeneity of the actors that need to agree to the policy change (Tsebelis
2002; König et al. 2010; Bräuninger et al. 2015). Therefore, fewer policy changes can be enacted
the greater the distance in preferences between the coalition partners. It is important to note that the
reduction of the set of feasible policy changes depends not on the number of coalition partners as
such, but on the relevant preference distance among them (Tsebelis 2002). In sum, coalition conflict
manifested in preference heterogeneity and polarization within a government decreases its policy-
making capacity and the increases the likelihood of legislative gridlock.
Coalition conflict (preference heterogeneity) is non-existent for single-party governments
under the usual assumption of parties as unitary actors. However, if the government party (or the
government coalition) does not command the necessary majority of votes in the legislature to pass
new policies and legislation (minority governments), the theoretically-relevant preference distance
might need to take into account other actors (for example, parties) whose consent would be needed,
even if formally these actors are not participating in government.
It should be noted that actual political systems differ in the way they organize the distribution
of responsibilities between the coalition partners and the relative autonomy the members of the
coalition enjoy within their sphere of responsibilities. But even where strong norms of ministerial
autonomy exist, the most important policies are decided by the cabinet collectively and need the
10
consent of the legislature, which brings back the relevance of the extent of coalition conflict for the
possibility of policy change (Andeweg 2000; Martin and Vanberg 2014). .
Greater coalition conflict and preference heterogeneity, more generally, have also been
hypothesized to increase the time needed for new policies to be agreed upon even when policy
change is feasible (Konig 2007; Rasmussen and Toshkov 2013; Martin and Vanberg 2004). Actors
who have more divergent preferences need more time and negotiation rounds to discover common
ground, reach a compromise, and arrange for the necessary side-deals that make a compromise
possible. According to Martin and Vanberg (2005, 2004), the incentive to scrutinize the coalition
partners increases with the ideological divergence within the coalition government, and
parliamentary scrutiny also takes time. In sum, we expect that:
H2: The greater the extent of government coalition conflict, (a) the lower the likelihood of policy
change and (b) the slower the policy change will occur.
The hypotheses above are rather standard in the literature on party government and have been tested
extensively before, although typically in terms of aggregate policy measures, like legislative
productivity and duration, rather than looking at concrete policy changes. What has been less
theorized about is the possibly moderating effect of coalition conflict on policy responsiveness, or
the link between public opinion and policy change. There are two aspects of this moderating effect.
One deals with the willingness and the other one with the capacity of coalition governments to
represent majority public opinion.
The capacity aspect was discussed above: to the extent that coalition conflict hampers
governments to enact policy changes, it would also hamper their capacity to enact the changes that
the median citizen supports. The flip side of this argument is that in times of high coalition conflict,
11
only the policy changes that enjoy very strong support by the public would have a chance. Gilens
(2012) explains how the difficulty of enacting policy change during times of high institutional
gridlock in the US means that mostly policies with strong political support get adopted. Moreover,
Coleman (1999) provides evidence that there is higher responsiveness to public opinion in periods
of unified compared to divided government in the US (see also Binder 1999).
The incentives aspect of the moderating effect of coalition conflict is not so straightforward.
Cabinets composed of more, and more diverse, parties are more likely to represent a broader range
of societal interests, and the (weighted) average of the policy positions of the parties in the coalition
would tend to approach the position of the median voter (Martin and Vanberg 2014). But to
hypothesize whether coalitions would be more or less likely to enhance policy responsiveness than
single-party governments, we have to take into account additional considerations. Theoretically,
single-party majority governments are strongly expected to reflect the position of the median voter,
provided that voters are policy motivated and there is a single relevant policy dimension. But a
broad coalition might actually be more representative than a single-party majority government in
the sense of being, on average, closer to the positions of the median voter when all policy
dimensions are considered. However, if individual coalition parties can control autonomously
particular policy domains, that might lead to more extreme policies being enacted than what the
median voter would like.
In addition, once in power, coalition governments might face weaker incentives to respond to
public opinion compared to single-party governments due to diffused responsibility (Duch et al.
2015). It is harder for people to judge the individual responsibility of parties part of a government
coalition and to attribute the blame when the government fails to adopt policies that the majority
favours. In turn, this decreases the pressure on individual parties to pander to public opinion, as the
political cost of non-responsiveness is shared or avoided. Altogether, we hypothesize that:
12
H3: The greater the extent of government coalition conflict, the weaker the positive effect of public
support for a policy alternative on (a) the likelihood of policy change and (b) the speed of policy
change.
These hypotheses take into account the differences between the preferences of the governing
parties, but not the preferences as such. In fact, we do not have many reasons to expect that the
general ideological preferences of parties should affect the overall likelihood of policy change,
unconditional on the content of the change. Only parties endorsing Conservatism as a political
ideology might be expected to hold a general bias in favour of the status quo due to their ideological
predispositions against social change, and then only if the traditional status quo has not already been
moved in a more progressive direction by their predecessors.
However, one might expect that governing parties would be more likely to enact specific
policy changes that agree with their general principles and party positions. In operational terms this
implies that policy change should be more likely to occur when it is in line with the aggregate party
positions of the governing parties. Despite the intuitive appeal of this reasoning, it has proved, in
fact, quite difficult to demonstrate empirically a link between the general positions of parties in
government and the content of policy changes they make (cf. the meta-analysis of Imbeau et al.
2001; and Schmitt 2016; as well as the more recent Knill et al. 2010). The contradictory empirical
evidence and operationalization difficulties notwithstanding, it remains theoretically plausible that
governing parties use their overall party positions as a shortcut and guide when it comes to taking a
stance on specific policy proposals in order to maintain a consistent platform and protect their brand
name vis-à-vis the electorate.
13
H4: Governments would be (a) more likely and (b) faster to enact changes that move policies in a
direction that is consistent with their general party positions.
These expectations should be even stronger when public opinion supports policy alternatives that
also coincide with the preferences of the parties in government. There is some empirical evidence
for such a conditioning effect of the fit between the direction of policy change and the preferences
of the government on the impact of public opinion (Petry and Mendelsohn 2004; Brettschneider
1996). For example, Brooks found that congruence on redistributive issues was higher for left-wing
than right-wing governments in France and the UK (1985, 1987). Moreover, a recent experiment in
Sweden provided evidence that party preferences may also constrain the extent to which individual
politicians respond to requests from citizens by demonstrating that politicians are less responsive to
requests which disagree with their party (Öhberg and Naurin 2016).
One of the reasons why policy responsiveness might not be perfect is precisely a clash of
public opinion with what the government parties want, either due to ideological commitments or the
need to cater to their own voters and sponsors (Petry and Mendelsohn 2004). In such cases, party
preferences may constrain responsiveness. But when a policy change is supported by the public and
it resonates positively with party preferences, there should be a very good chance of the policy
being enacted. Moreover, the combined effect should be more than the sum of its parts. Assuming a
heterogeneous set of policy alternatives, some of which require only popular support, some only
party support, and some both popular and party support to get enacted, we arrive at theoretical
expectations of interaction effects between public and party support:
14
H5: The positive effect of public support on (a) the likelihood and (b) speed of policy change should
be stronger for policy changes that move policies in a direction consistent with the general party
positions of the governments.
To sum up the theoretical discussion so far, we hypothesize direct effects of public support,
coalition conflict, and government support for policy change on the likelihood and speed of policy
change, as well as interaction effects between public opinion on the one hand, and coalition conflict,
and government support for policy change, on the other hand. In the next section we explain how
we design the empirical tests of these hypotheses.
Research design, data and operationalization
In general terms, our research design is based on a comparison of the speed and occurrence of
policy change on a large number of policy issues across several national political systems. The
countries we study – Denmark, Germany, and the UK – are all established parliamentary
democracies that differ, however, in their characteristic patterns of party government. While single-
party majority cabinets are common in the UK, multiparty coalitions are typical in Germany and in
Denmark, where one also observes the phenomenon of multi-party minority coalition cabinets.
During the time-period of our analysis, the extent of actual coalition conflict differed not only
between the three countries, but also within the countries over time. For example, the UK
experienced both a single-party majority cabinet and a two-party coalition, Germany went through
several coalitions that varied in the range of preferences of the participating parties, and Denmark
went through a number of successive coalitions, both majority and minority ones.
15
Unit of observation and sample selection
Our unit of observation is a policy issue in a country over time, and we analyse a total of 306 issues
in the three countries. We look at concrete policy issues, rather than aggregate policy output
measures or latent policy dimensions. For each policy issue, we identify public support for a policy
alternative (call for public action) that relates to specific measures that the national politicians can
enact. In line with existing research (Gilens and Page 2014; Gilens 2012) we then follow each of
these issues from the time of the public opinion poll until the policy change is enacted or, if that
does not happen, to a maximum of 48 months. Our focus on concrete issues has the advantage that
we relate public opinion and public policy directly (Gilens 2012; Lax and Phillips 2009; Burstein
2014). By selecting a time window of up to four years, we allow ample time for new proposals to
enter the legislative agenda and get adopted.
To select the policy issues we analyse, we started with identifying relevant questions asked in
representative nationwide public opinion surveys in Denmark (1998-2010), Germany (1998-2010)
and the UK (2001-2010). To be relevant, the questions had to tap into the attitudes of the adult
population towards issues of public policy, to involve a call for future political action, and to relate
to specific policy issues. In addition, the questions had to concern issues of national policy
competences, and the responses had to be measured on a scale on which respondents expressed the
extent to which they agreed or not with a given policy change.
We identified a total of 102 survey questions that fulfilled these criteria in Germany, 211 in
Denmark, and 239 in the UK3. In our final sample, we took all the 102 relevant questions in
3 In Denmark, all selected survey questions came from surveys conducted by the Gallup Institute. In Germany, we
relied on questions from asked by the Politbarometer surveys. For the United Kingdom, we relied on a list of questions
16
Germany, and we used random sampling stratified by year to select 102 questions from Denmark
and the UK each, for a total of 306 cases4.
In all three countries the selected survey questions cover a wide range of different policy
issues and relate to different policy areas that represent diverse policy types (regulatory,
(re)distributive and constituent policies). For instance, in the UK the sample includes questions
concerning a possible amnesty to illegal immigrants, the introduction of an identity card system,
and the replacement of university tuition fees with a graduate tax scheme.
Our sampling strategy is constrained by the availability of reliable public opinion data
representative at the national level (Gilens 2012; Monroe 1998). As a consequence, our sample
might be biased towards more salient issues that are more likely to get the attention of polling
companies. The potential bias is not necessarily a problem since focusing on questions that have at
least some amount of salience makes it “plausible that average citizens may have real opinions and
may exert some political influence” (Gilens and Page 2014: 568). Still, it is worth reminding that
our sample of issues might not be representative of the universe of all possible policy issues of
national competence that could have been on the agenda5, despite the fact that it covers a broad
range of policies in terms of type and domain.
from YouGov and ICM sampled by XX, which was further appended by additional survey questions from the
mentioned companies
4 The sampling was necessitated by the high costs of data collection per policy issue and to a lesser extent by the need to
keep the number of cases balanced across the three countries.
5 A recent US study on responsiveness addresses this challenge by sampling its cases from proposals on the legislative
agenda rather than from available polls (Burstein 2014). However, constructing such a potential universe of issues is
difficult in the context of a cross-country comparative study like ours, given that there is no comparative sampling
frame of all possible issues that would be applicable to all three countries. Moreover, relying on national legislative
17
To further address the potential issue of non-representativeness, we measure and control for
the media salience of all issues in our sample, and we note and control for whether policy change
was already on the legislative agenda at the time of the public opinion poll. On salient issues it
might be easier for policy makers to acquire information about public views and more costly to
depart from them (Lax and Phillips 2012). Our sample covers issues of rather different salience
within each of the three countries (with some issues not receiving any coverage in our newspapers
at all), and only 51 of the 306 issues were related to an existing bill proposal or cabinet decision
when the opinion question was asked. In sum, we can be confident that our sample includes issues
of varying media salience and at various stages in the policy-making process.
Outcome variables
In the empirical analyses we study both the occurrence and speed of policy change, as two
theoretically-relevant aspects of public policy making. In the first set of analyses we examine the
likelihood of policy change (i.e. whether the national government or parliament adopted primary or
secondary legislation in line with the public call for change) for each government that was in office
within the observation period that we follow the policy debate on a given policy issue. In the second
set of analyses we focus on the duration between the date of the public opinion survey and the date
of the policy change, if it occurred within the four-year case-specific period of observation on the
306 issues. To detect the occurrence and timing of policy change we relied on historic information
databases or media in the three countries would only yield information about issues that have passed a first “threshold
of access” by being picked up by either politicians or the media, which might create another source of bias.
18
provided by legislative databases, other government (web)sources, online newspaper archives, and
information provided by interest groups and professional associations6.
Explanatory variables
We operationalize public support for policy change as the percentage of all respondents in favour of
the call for policy action as expressed in the public opinion survey. To further explore the possible
effect of public opinion, we employ three additional operationalizations – majority public support (a
dichotomous variable indicating whether the call for policy action enjoys the support of a majority
of all respondents), net public support (a variable with a theoretical range between -1 and +1,
defined as the share of respondents in favour of the call for policy action minus the share opposed),
and public support calculated as the share of respondents in favour from those with an opinion
(hence, excluding no responses and ‘don’t knows’)7.
The measures of coalition conflict and government support for policy changes require that we
obtain estimates of relevant government and party positions. To do that, we use the Chapel Hill
expert survey of party positions (Bakker et al. 2015) 8. We make the measures policy scale-specific:
6 Descriptive statistics of the variables used in the analysis are reported in the Supplementary Material (Table A1).,
available as an online appendix to this paper.
7 The statistical models employing these alternative measures of public opinion are reported in Table A2 in the
Supplementary Material.
8 We considered using the Manifestos Project data (Klingemann et al. 2007) as an alternative source of party and
government positions. We obtained the necessary data and constructed measures on 10 scales constructed from the
Manifesto items related to our policy issues. The data, however, failed face validity checks as it provided implausible
estimates of party positions and relative ranking of the parties on scales and positions (details available upon request).
19
first we classify each of the 306 policy issues to one of the three main scales (or dimensions) in the
Chapel Hill dataset – general left/right, economic left/right and GAL-TAN (green, alternative, and
liberal vs. traditional, authoritarian and nationalist), then we identify the relevant party positions on
these scales and assign them to the case, and finally we compute the coalition conflict and
government support measures.
Coalition conflict is operationalized as the absolute distance between the positions of the two
most extreme government parties (Tsebelis 2002; Hartmann 2015; König et al. 2010) on the
relevant policy scale9. Government positions are operationalized as the weighted average of the
positions of the parties in government, with the weights corresponding to the relative shares of the
seats in the legislature held by the coalition party from the total number of seats held in the
legislature by all coalition partners10. To obtain a measure of relative government support for policy
change from absolute government positions, we have to consider the direction of policy change for
each case. To that end, we first code the implied direction of each policy change (e.g. left or right),
Hence, we decided against reporting results based on this data source, which is more appropriate for measuring party
attention to particular issues rather than positions as such.
9 The Danish case presents a theoretical challenge for the measurement of average government positions and coalition
conflict because of the minority status of the cabinets in the country during our period of observation. Taking into
account only the parties formally part of the (minority) coalition might underestimate the degree of intra-government
conflict and misrepresent the average government position since the governing parties need the support of additional
parties in the legislature to pass legislation. At the same time, minority coalitions have flexibility in choosing a partner
in the legislature for particular policy proposals that is not easily captured. Nevertheless, we constructed alternative
measures of government positions and coalition conflict in Denmark that take into account the unofficial but regular
legislative partners of the parties in the governing minority coalitions. The results based on these alternative measures
can be found in Table A2 in the Supplementary Material.
10 For easier interpretation, the original scales, which range from 1 to 10, are centered at zero.
20
and then we invert the original government positions where needed, so that more right wing parties
are aligned with right-leaning policy changes and left wing parties with left-leaning changes11.
In addition to these main variables of interest, we include a variable that indicates whether the
call for public action was related to an existing bill proposal or a cabinet decision when the public
opinion question was asked, because such cases could have a higher likelihood of policy change.
Also, we measure the media saliency of each case by tracking the number of newspaper articles
related to the case in one major national newspaper in each country (Politiken in Denmark,
Sueddeutsche Zeitung in Germany, and The Guardian in the UK) in the period between one month
prior and one month after the public opinion survey was conducted.12 All empirical models reported
below include country fixed effects (dummies) in order to control for unobserved country-level
heterogeneity in the likelihood of occurrence of policy change and in the duration until change
occurs. Because in the logistic regression models the unit of analysis is a government spell (a period
of time during which a government is responsible for a policy issue), we also include as a control
the remaining formal tenure of the government (in months) from the moment of its inauguration or
11 For example, if the call for action concerned increasing taxation (left-leaning policy change), we inverted the position
scores of the relevant government, so that higher scores would be associated with more economically left-wing
positions (on the original scales, higher scores are associated with more right/TAN positions).
12 The Boolean media keyword search was conducted using the FACTIVA database. A complete list, including all
Boolean search requests, for all the 306 survey questions will be provided by the authors on request. The keywords
represent the respective items as accurately as possible while paying attention to the scope of the policy item. We
included the plurals of the selected key words, their word stems, and their synonyms. The media count for each country
is standardized, i.e. the variable is rescaled to have a mean of zero and a standard deviation of one in each country. By
measuring issue coverage within a fixed two month period, we avoid bias resulting from the fact that opinion items
which experience a policy change receive higher media attention in the time period before the actual policy change
occurs.
21
the date of the public opinion survey (whichever comes last) to the moment of its expected
dissolution or the end of our observation period (whichever comes first).
Empirical analyses
Public opinion and policy change
We present a detailed analysis of the bivariate relationships between public opinion and policy
change in the Supplementary Material (Figures A1 and A2 and the associated text). In summary, in
all three countries the likelihood of policy change increases with higher levels of public support,
both absolute and net. In substantive terms, however, the effect of public opinion is relatively small.
We also observe rather modest levels of congruence between policy and majority public opinion at
the beginning of the observation period, and even more modest levels of improvements in
congruence over time, despite a considerable degree of policy-making activity13.
Multivariate logistic regression models
Table 1 presents the results of a series of multivariate logistic regression models of the likelihood of
policy change within a period of four years following the polling of public opinion on a call for
policy action for each government that was in office within the observation period. Model 1
includes the main variables of interest and the controls, but no interactions. Model 2 adds the
13We should not that the positive effect of public opinion on the likelihood on policy change does not appear to be
linear, but the exact form of the relationship differs across the three countries and does not follow a simple interpretable
pattern (for details, see Figure A2 in the Supplementary Material). The non-linearity is less pronounced in a multivariate
setting, however.
22
interaction between coalition conflict and public opinion. Model 3 adds the interaction between
government support and public support.
The effect of public support is consistently positive, and it is statistically significant in Models
1 and 3 at the 5% level. In substantive terms the effect is relatively small: for an increase of public
support for a policy from 50% to 60%, the coefficients from Model 1 imply that the likelihood of
policy change increases with 2.2 percentage points (other covariates held constant at their means or
typical values).
Table 1. Logistic regression models of policy change
Model 1 Model 2 Model 3
Public support 1.44 (0.56) * 1.30 (0.89) 1.48 (0.56) **
Coalition conflict -0.02 (0.10) -0.02 (0.09) -0.02 (0.10)
Public support*Coal. conflict / 0.08 (0.43) /
Government support 0.15 (0.09) 0.15 (0.09) 0.16 (0.09)`
Gov. support* Public opinion / / -0.32 (0.42)
Media saliency 0.32 (0.10) ** 0.32 (0.10) ** 0.33 (0.10) **
Existing proposal 0.95 (0.29) ** 0.95 (0.29) ** 0.96 (0.29) **
Remaining months 0.04 (0.01) *** 0.04 (0.01) *** 0.04 (0.01) ***
Denmark 0.35 (0.32) 0.35 (0.32) 0.33 (0.32)
United Kingdom -0.92 (0.35) ** -0.92 (0.35) ** -0.92 (0.35) **
Intercept -3.02 (0.50) *** -3.02 (050) *** -3.02 (0.50) ***
AIC 516.31 518.27 517.75
Notes: Logistic regression models (with logit link). Dependent variable: occurrence of policy change. Unit of analysis is
a government spel; N=525. Unstandardized and unexponentiated coefficients. Public opinion centered at 0.5.
Government positions centered at 5.0. Media saliency is logged. Significance levels: 0 < *** < 0.001 < ** < 0.01 * <
0.05 ` < 0.1
23
The analysis provides no evidence that policy change is more likely under governments composed
of parties with relatively similar positions or under single-party governments. Coalition conflict has
no significant effect across the model specifications, which, to remind, include country fixed effects
(see also Figure A2 in the Supplementary Material). According to Model 2, there is also no
interaction between public support and coalition conflict. Inspecting the interaction visually
confirms this inference and shows that even when coalition conflict is set at the minimum and
maximum observed values, the confidence intervals for the effect of public opinion under these two
scenarios overlap for the entire range of public opinion (see Figure A3)14.
Government support for the policy proposal has a positive effect that is close to statistical
significance (the p-value of the effect in the three models ranges between 0.09 and 0.11). According
to Model 315, the interaction between public and government support is, contrary to expectation,
negative, although non-significant at the 95% level. When we plot the interaction, we can see that
the sensitivity of the likelihood of policy change on the level of public support is stronger for policy
changes that lack government support and is rather flat for policies that enjoy government support.
Similarly, the effect of government support on the likelihood of policy change is rather steep for
policy changes that lack public support, but non-existent for polices that enjoy public support
(Figure A5 illustrating these effects is to be found in the Supplementary Material). Overall, these
findings do not present evidence that policy responsiveness to public opinion is less likely on policy
14 When we use the alternative operationalizations of coalition conflict and government support that take into account
the unofficial partners of the governing parties in the minority Danish cabinets, we find essentially the same results.
15 In Table A2 in the Supplementary Material we report versions of Model 3 that include robust (A7) or clustered
standard errors (A8) and a multi-level specification with the policy issue (case) as a second level of analysis (A6). The
inferences remain essentially the same.
24
issues that would move policy in a direction opposite to the general party positions of the
government16. If anything, public support for a policy seems to matter more when the policy
changes are not in line with the preferences of the government. If we take the results of Model 3 and
Figure A5 seriously despite the lack of formal statistical significance, it would appear that a policy
change has a very similar chance of being adopted (a) under a supportive government irrespective
of its level of public support and (b) under an opposing government but only if it enjoys very high
levels of public support.
Finally, we should note that the three control variables have the expected effects in Models 1-
3: the formal time of a government remaining in office, the prior existence of a government bill and
media salience all increase the likelihood of policy change. The existence of a bill makes it more
than 2.5 more likely that policy change will follow. Doubling the number of newspaper articles on a
topic (salience) is associated with a 40% higher risk of policy change, and each additional month in
office adds approximately a 4% increase in the odds. But importantly in light of existing literature
that has emphasized the potential role of salience in moderating the impact of public opinion on
policy change, there is no evidence for an interaction between salience and public support in our
dataset (model not shown).
16 Interestingly, while government support for policy change as such has a rather weak and only marginally significant
positive effect, the absolute position of the governments appears to be negatively and significantly associated with the
likelihood of policy change. That is, more right and nationalist (TAN) governments area associated with a lower
likelihood of policy change, irrespective of the implied direction of policy change. In addition, they appear to be less
responsive to public opinion as well: there is some marginal evidence for a negative interaction between public support
and government positions (for details, see Model A5 in Table A2 in the Supplementary Material).
25
Event history models of policy-making duration
The analyses so far focused on the question whether policy change occurred at all within a four-year
period after the public opinion survey was conducted. The speed of policy change, however, is
another important aspect that deserves attention in its own right17.
Table 2 presents the results of three Cox proportional hazards models with time-varying
measures of government positions and coalition conflict18. Model 4 has the main effects of interest,
while Models 5 and 6 add the two interactions – between coalition conflict and public support and
between government support and public support, respectively. The findings are very similar to the
logistic regressions of the likelihood of policy change presented above. Public support has a
consistently positive and significant (in Models 4 and 6) effect on the hazard of policy change
(hence, it is associated with shorter durations until change occurs). Coalition conflict has no effect,
while the main effect of government support is positive and marginally significant leaving some
evidence that policy changes in line with the general positions of the government tend to occur
faster.
Similarly to Model 3, Model 6 reports a negative interaction between government and public
support, which is however not significant. In any case, given that our hypothesis predicted a
positive interaction, it is safe to say that the empirical analysis does not find support for this
hypothesis.
17 In addition, the event history analysis can address the complication that for many of our cases the observation time is
censored at four years, which of the length of time we followed each policy issue in line with the approach of Gilens
(2012) and Gilens and Page (2014).
18 We explore in more detail the impact of the individual variables on the timing of policy change in Figures A6, which
presents the Kaplan-Meier survival curves, and the associated text in Supplementary Material.
26
Table 2. Cox Proportional Hazards models with time-varying covariates
Model 4 Model 5 Model 6
Public support 0.95 (0.46) * 0.78 (0.78) 1.04 (0.45) *
Coalition conflict 0.03 (0.09) 0.02 (0.09) 0.02 (0.09)
Government support 0.11 (0.07)’ 0.11 (0.07)’ 0.13 (0.07)’
Public support*Coalition conflict / 0.11 (0.38) /
Public support*Government support / / -0.32 (0.31)
Media saliency 0.26 (0.09) ** 0.26 (0.09) ** 0.27 (0.09) **
Existing proposal 0.79 (0.22)*** 0.79 (0.22)*** 0.81 (0.21) ***
Notes: Cox Proportional Hazards models with time-varying covariates. Dependent variable: occurrence of policy
change on a policy in a month. N=10815. The models are stratified by country and clustered by case id. Unstandardized
and unexponentiated coefficients. Public opinion centered at 0.5. Media saliency is logged. Significance levels: 0 < ***
< 0.001 < ** < 0.01 * < 0.05 ` < 0.1
Conclusion
Recent years have witnessed an expansion of the study of responsiveness to a broad range of
political systems and a new research agenda has started exploring how contextual differences in
institutional architectures might contribute to explaining varying degrees of opinion-policy linkage
(Wlezien and Soroka 2012; Lax and Phillips 2012; see e.g. Hobolt and Klemmensen 2008). At the
same time, other contextual factors that may affect responsiveness, such as the impact of cabinet
politics and coalition governance, have received less attention. By linking literature on political
responsiveness, party government and coalition conflict, we identified a number of hypotheses how
government positions and coalition conflict could not only impact policy making directly, but also
moderate the degree of policy responsiveness.
27
We found a systematic but substantially rather small degree of policy responsiveness in the
three parliamentary political systems examined. In Germany, despite a considerable amount of
policy activity on the set of cases, the probability of policy change was not strongly affected by the
degree of public support and, as a result, overall congruence between the wishes of the majority and
the state of policy hardly improved over a four-year observation period. Responsiveness was higher
in the case of the UK, but coupled with a rather stronger status quo bias of the British policy-
making system this also did not produce a high degree of congruence between policy and majority
public opinion. In Denmark, moderate responsiveness and relatively high degree of policy-making
activity produced the highest degrees of congruence we observed, although in absolute terms
congruence was still disappointingly low.
The fact that Denmark – governed by minority coalitions throughout our study period –
exhibited the highest ability to produce policy change, while the UK – governed for a large part of
the observation period by a single party majority government – experience the least amount of
policy change already spelled trouble for the hypothesis that the extent of coalition conflict
significantly affects policy-making capacity and, indirectly, responsiveness. Accordingly, we found
no evidence that coalition conflict has a negative effect on policy adoption or that it moderates
policy responsiveness. The lack of an effect of coalition conflict19 on policy-making capacity is
unexpected, although in hindsight we can evoke reference to the concept of ministerial autonomy to
19 In line with existing literature, our measure of coalition conflict is based on the maximum distance between the
coalition partners in a government and, as such, disregards the relative power of the different coalition partners. It could
be that a measure that takes into account not only the preference distance but the distribution of power as well would
capture better the theoretical concept. The construction of such a measure is, however, nontrivial as there are multiple
ways in which preference distance and power distribution can be combined to produce a single measure.
28
rationalize the null result (Martin and Vanberg 2014; e.g. Laver and Shepsle 1994). And perhaps the
fact that coalition government are more likely to represent the societal median (Blais and Bodet
2006; Powell and Vanberg 2000) compensates for their allegedly diminished capacity for policy
change, so that responsiveness remains similar overall. Moreover, it needs to be remembered that
even if holding governments to account may more difficult in countries with frequent coalition
governments, coalition governance does not only make it harder for governments to act on popular
policies, but may also prevent them from adopting unpopular ones in practice (Gilens 2012). If
these two effects counterbalance each other, we would not expect coalition conflict to affect
responsiveness.
We also found no support for the hypothesis that there should be a positive interaction
between government support for policy change and public opinion, when they coincide. This
implies that responsiveness is not constrained when the policy proposals on the agenda run counter
to the general policy positions of the parties. We did not find evidence that responsiveness is
weaker on issues that move policy in the opposite direction to the general preferences of the
government. If anything, it would appear that government and public support are substitutes, so that
public support for policy change matters more when the policy does not enjoy government support,
and government support for policy change matters more when the policy change is not favoured by
the public. Instead of strong effects of relative government support, our analyses hint that the
absolute government positions might have an independent impact on the probability of policy
change. Governments with positions located towards the right/TAN ends of the relevant policy
scales appeared to be less likely to enact policy changes and less responsive to public opinion.
Altogether, we do not find much evidence that patterns of party government in Western
Europe weaken policy responsiveness. Moreover, while responsiveness in our examined countries
is rather low in absolute terms, neither coalition heterogeneity nor discrepancy between an issue and
29
the aggregate policy positions of the parties lower the likelihood that governments respond to public
opinion. Future research should extend our study of responsiveness to other parliamentary systems
with strong parties. While it is reassuring to also find signs of responsiveness in such systems, the
complex relationships between coalition conflict, partisan preferences, and responsiveness in these
contexts deserve further scrutiny.
A promising approach for such research would be to use direct measures of party preferences
related to specific issues. Our study already improves on existing literature by linking public policy,
public opinion and government positions on three different scales, rather than simply using general
left-right ideological positions as proxies for government preferences. But there could be
advantages of looking into even more concrete party positions on policy issues in order to
disentangle the effects of government preferences on policy changes and responsiveness.
There is also scope for extending our research to studies of the dynamic relationship between
opinion and policy over time in Western European parliamentary democracies. As mentioned,
responsiveness is likely to be a reciprocal relationship, in which both opinion and policy adapt to
each other. In a study of a high number of different issues like ours, examining the dynamic
relationship is not possible due to the limited availability of repeated opinion polls on the same
specific topic in the examined countries. By focusing on specific issues, we address one of the
criticisms of studies linking general measures of opinion and policy when it comes to assessing
causality. They face the potential challenge that the issues used to construct the aggregate opinion
and policy measures may not be the same (Lax and Phillips 2012; Burstein 2014). Instead, our
approach gives us confidence that the public has expressed its attitudes towards the same policies as
the ones for which we measure policy outcomes. However, future research should complement our
research by scrutinizing the reciprocal linkage between opinion and policy further in studies of the
30
small subset of specific policy issues for which time series data is available and by relying on
qualitative and experimental methods.
31
References
Achen, Christopher H., and Larry M Bartels. 2016. Democracy for Realists: Why Elections
Do Not Produce Responsive Government. Princeton: Princeton University Press.
Andeweg, Rudy B. 2000. "Ministers as double agents? The delegation process between
cabinet and ministers." European Journal of Political Research 37 (3):377-95.
Bakker, Ryan, Catherine de Vries, Erica Edwards, Liesbet Hooghe, Seth Jolly, Gary Marks,
Jonathan Polk, Jan Rovny, Marco Steenbergen, and Milada Anna Vachudova. 2015.
"Measuring party positions in Europe: The Chapel Hill expert survey trend file, 1999–
2010." Party Politics 21 (1):143-52.
Bardi, Luciano, Stefano Bartolini, and Alexander H. Trechsel. 2014. "Responsive and
Responsible? The Role of Parties in Twenty-First Century Politics." West European
Politics 37 (2):235-52.
Bevan, Shaun, and Will Jennings. 2014. "Representation, agendas and institutions." European
Journal of Political Research 53 (1):37–56.
Binder, Sarah A. 1999. "The Dynamics of Legislative Gridlock, 1947-96." American Political
Science Review 93 (3):519-33.
Blais, André, and Marc André Bodet. 2006. "Does Proportional Representation Foster Closer
Congruence Between Citizens and Policy Makers?" Comparative Political Studies 39
(10):1243-62.
Brettschneider, Frank. 1996. "Public opinion and parliamentary action: responsiveness of the
German Bundestag in comparative perspective." International Journal of Public
Opinion Research 8 (3):292-311.
Brooks, Clem, and Jeff Manza. 2006. "Social policy responsiveness in developed
democracies." American Sociological Review 71 (3):474-94.
32
Brooks, Joel E. 1985. "Democratic Frustration in the Anglo-American Polities: a
Quantification of Inconsistency Between Mass Public Opinion and Public Policy."
Western Political Quarterly 38 (2):250-61.
———. 1987. "The Opinion-Policy Nexus in France: Do Institutions and Ideology Make a
Difference?" The Journal of Politics 49 (2):465-80.
Bräuninger, Thomas, Marc Debus, and Fabian Wüst. 2015. "Governments, Parliaments and
Legislative Activity." Political Science Research and Methods 3 (1):1-26.
Burstein, P. 2014. American public opinion, advocacy and policy in Congress. Cambridge:
Cambridge University Press.
Coleman, John J. 1999. "Unified Government, Divided Government, and Party
Responsiveness." The American Political Science Review 93 (4):821-35.
Coman, Emanuel Emil. 2015. "Electoral proportionality, multi-party cabinets and policy
responsiveness." Electoral Studies 40:200-9.
Duch, Raymond, Wojtek Przepiorka, and Randolph Stevenson. 2015. "Responsibility
Attribution for Collective Decision Makers." American Journal of Political Science 59
(2):372-89.
Erikson, Robert S. 2015. "Policy Responsiveness to Public Opinion." In Oxford
Bibliographies. DOI: 10.1093/OBO/9780199756223-0103: DOI:
10.1093/OBO/9780199756223-0103.
Erikson, Robert S., Michael B. Mackuen, and James A. Stimson. 2002. The Macro Polity. .
Cambridge: Cambridge University Press.
Gilens, Martin. 2012. Affluence and influence: Economic inequality and political power in
America. Princeton: Princeton University Press.
33
Gilens, Martin, and Benjamin I. Page. 2014. "Testing Theories of American Politics: Elites,
Interest Groups, and Average Citizens." Perspectives on Politics 12 (3):564-81.
Hartmann, Sebastian. 2015. Partisan Policy-Making in Western Europe: Springer.
Hobolt, Sara Binzer, and Robert Klemmensen. 2008. "Government Responsiveness and
Political Competition in Comparative Perspective." Comparative Political Studies 41
(3):309-37.
Imbeau, Louis M., François Pétry, and Moktar Lamari. 2001. "Left–right party ideology and
government policies: A meta–analysis." European Journal of Political Research 40
(1):1-29.
Jennings, Will, and Peter John. 2009. "The Dynamics of Political Attention: Public Opinion
and the Queen's Speech in the United Kingdom." American Journal of Political
Science 53 (4):838-54.
Klingemann, Hans-Dieter, Andrea Volkens, Judith Bara, and Ian Budge. 2007. Mapping
Policy Preferences II: Estimates for Parties, Electors and Governments in Central and
Eastern Europe, European Union and OECD 1990-2003. Oxford: Oxford University
Press.
Knill, Christoph, Marc Debus, and Stephan Heichel. 2010. "Do parties matter in
internationalised policy areas? The impact of political parties on environmental policy
outputs in 18 OECD countries, 1970–2000." European Journal of Political Research
49 (3):301-36.
Konig, Thomas. 2007. "Divergence or convergence? From ever-growing to ever-slowing
European legislative decision making." European Journal of Political Research 46
(3):417-44.
34
König, Thomas, George Tsebelis, and Marc Debus. 2010. Reform processes and policy
change: veto players and decision-making in modern democracies. Vol. 16: Springer
Science & Business Media.
Laver, Michael, and Kenneth A Shepsle. 1994. "Cabinet ministers and parliamentary
government." Cambridge: Cambridge University Press.
Lax, Jeffrey R., and J. H. Phillips. 2009. "How Should We Estimate Public Opinion in The
States ?" American Journal of Political Science 53 (1):107-21.
Lax, Jeffrey R., and Justin H. Phillips. 2012. "The Democratic Deficit in the States."
American Journal of Political Science 56 (1):148-66.
Martin, Lanny W., and Georg Vanberg. 2004. "Policing the Bargain: Coalition Government
and Parliamentary Scrutiny." American Journal of Political Science 48 (1):13-27.
———. 2005. "Coalition Policymaking and Legislative Review." American Political Science
Review 99 (1):93-106.
———. 2011. Parliaments and Coalitions. The Role of Legislative Institutions in Multiparty
Governance. New York: Oxford University Press.
———. 2014. "Parties and Policymaking in Multiparty Governments: The Legislative
Median, Ministerial Autonomy, and the Coalition Compromise." American Journal of
Political Science 58 (4):979-96.
McDonald, Michael D, Silvia M Mendes, and Ian Budge. 2004. "What are elections for?
Conferring the median mandate." British Journal of Political Science 34 (01):1-26.
Monroe, Alan D. 1998. "Public Opinion and Public Policy, 1980-1993." The Public Opinion
Quarterly 62 (1):6-28.
35
Müller, Wolfgang C. 2000. "Political parties in parliamentary democracies: Making
delegation and accountability work." European Journal of Political Research 37
(3):309-33.
Page, B. I., and Robert Y. Shapiro. 1983. "EFFECTS OF PUBLIC-OPINION ON POLICY."
American Political Science Review 77 (1):175-90.
Petry, François, and Matthew Mendelsohn. 2004. "Public opinion and policy making in
Canada 1994-2001." Canadian Journal of Political Science-Revue Canadienne De
Science Politique 37 (3):505-29.
Powell, G Bingham. 2006. "Election laws and representative governments: Beyond votes and
seats." British Journal of Political Science 36 (02):291-315.
Powell, G Bingham, and Georg S Vanberg. 2000. "Election laws, disproportionality and
median correspondence: Implications for two visions of democracy." British Journal
of Political Science 30 (03):383-411.
Przeworski, Adam. 2010. Democracy and the Limits of Self-Government. Cambridge:
Cambridge University Press.
Rasmussen, Anne, and Dimiter Toshkov. 2013. "The effect of stakeholder involvement on
legislative duration: Consultation of external actors and legislative duration in the
European Union." European Union Politics 14 (3):366-87.
Saeki, Manabu. 2009. "Gridlock in the Government of the United States: Influence of Divided
Government and Veto Players." British Journal of Political Science 39 (03):587-607.
Schermann, Katrin, and Laurenz Ennser-Jedenastik. 2014. "Coalition Policy-Making under
Constraints: Examining the Role of Preferences and Institutions." West European
Politics 37 (3):564-83.
36
Schmitt, Carina. 2016. "Panel data analysis and partisan variables: how periodization does
influence partisan effects." Journal of European Public Policy 23 (10):1442-59.
Shapiro, Robert Y. 2011. "Public Opinion and American Democracy." Public Opinion
Quarterly 75 (5):982-1017.
Soroka, Stuart, and Christopher Wlezien. 2010. Degrees of Democracy: Politics, Public
Opinion, and Policy. Cambridge: Cambridge University Press.
Strøm, Kaare, Wolfgang C. Müller, and Daniel Markham Smith. 2010. "Parliamentary
Control of Coalition Governments." Annual Review of Political Science 13 (1):517-35.
Toshkov, Dimiter. 2011. "Public opinion and policy output in the European Union: A lost
relationship." European Union Politics 12 (2):169-91.
Tsebelis, G. 1999. "Veto players and law production in parliamentary democracies: An
empirical analysis." American Political Science Review 93 (3):591-608.
Tsebelis, George. 2002. Veto Players: How Political Institutions Work. Princeton: Princeton
University Press.
Weissberg, Robert. 2002. Polling, Policy, and Public Opinion: The Case Against Heeding
the" voice of the People": Palgrave Macmillan.
Wlezien, Christopher. 1995. "The Public as Thermostat: Dynamics of Preferences for
Spending." American Journal of Political Science 39 (4):981-1000.
———. 2016. "Public Opinion and Policy Representation: On Conceptualization,
Measurement, and Interpretation." Policy Studies Journal:n/a-n/a.
Wlezien, Christopher, and Stuart Soroka. 2016. "Public Opinion and Public Policy." In
Oxford Research Encyclopedia. Oxford: Oxford University Press.
Wlezien, Christopher, and Stuart N. Soroka. 2012. "Political Institutions and the Opinion–
Policy Link." West European Politics 35 (6):1407-32.
37
Öhberg, Patrik, and Elin Naurin. 2016. "Party-constrained Policy Responsiveness: A Survey
Experiment on Politicians’ Response to Citizen-initiated Contacts." British Journal of
Political Science 46 (4):785-97.
38
Appendix (Supplementary Material)
Contents
1. Descriptive statistics (Table A1)
2. Public opinion and policy change: descriptive and bivariate analyses (Figures A1 and
A2)
3. Coalition conflict and policy change and responsiveness (Figures A3 and A4)
4. Additional logistic regression models of policy change (Table A2)
5. Additional figures illustrating the interaction effects (Figures A5)
6. Time until policy change: bivariate analysis (Figure A6)
1. Descriptive statistics
Table A1. Descriptive statistics
Minimum Mean Median Maximum St. dev.
Public support 0.04 0.51 0.52 0.97 0.21
Net public support -0.90 0.12 0.17 0.96 0.40
Media salience 0.00 10.66 4.00 153.00 18.38
Duration of policy change (in months);
censored at 48 months 0.00 14.97 10.00 46.00 13.54
Coalition conflict 0.00 1.47 0.94 4.09 1.32
Government preferences 3.27 5.33 5.07 7.63 1.22
Distribution
Existing proposal [yes/no] 17%/83%
Direction of policy change [right/left] 56%/44%
Policy change [yes/no] 39%/61%
Policy scale [general LR/econ./GALTAN] 38%/45%/18%
39
2. Public opinion and policy change: Descriptive and bivariate analyses
Figure A1. Public opinion and policy change in Western Europe: responsiveness and congruence
The top two panels of Figure A1 show how the likelihood of policy change varies with the
percentage support for policy change (top-left) and with net public support for policy change
(top-right). The lines for each of the three countries are based on the estimates from country-
level logistic regression models of policy change regressed on public support and net public
support, respectively. Clearly, in all three countries the likelihood of policy change increases
with higher levels of public support, both absolute and net.
40
The strength of responsiveness is greatest in the UK and smallest in Germany. In
substantive terms, however, the effect of public opinion is relatively small – for a change in
the share of the public supporting policy change from 0 to 1 (a rather unlikely change), the
likelihood of the government adopting the policy only shifts from 37% to 54% in Germany,
from 34% to 61% in Denmark, and from 10% to 30% in the UK. For an increase in public
support of 10 percentage points (still large, but more realistic), the likelihood of policy change
increases with around 1.6 percentage points in Germany, around 2.6 percentage points in
Denmark, and with around 3 percentage points in the UK. Also, note that the lines cross the y-
axis at rather high points meaning that policy change has a substantial chance of happening
even in the complete lack of public support.
The positive relationship between public support and the likelihood of policy change is
not statistically significant in the country-level models without additional covariates. It is also
worth noting that a policy change needs to enjoy at least 60% public support in Denmark and
80% in Germany to have a 50% or higher chance of being enacted within the next four years.
In the UK sample, the bias towards the status quo is even more pronounced with policy
change having an estimated maximum likelihood of only 42% even with maximum public
support.
Another way to explore the relationship between public opinion and policy change is to
examine the percentage of policies that are congruent with majority public opinion (meaning
that the policy status quo at the time has the support of the majority of the public). The
bottom-left panel of Figure A1 tracks average congruence for each country over our four-year
period of observation. The three countries start with similar levels of congruence at the time
when the opinions polls are taken (with public support favouring the status quo in 41% in
Denmark and the UK and 48% in Germany). Over time, congruence rises to 50% in the UK,
to 51% in Germany, and to 61% in Denmark. In sum, for the set of policy issues in our
sample, four years after the initial call for public action congruence with public opinion is
present in around half of the cases in Germany and the UK and round 3 out of 5 cases in
Denmark.
These very modest levels of congruence between policy and majority public opinion
and even more modest levels of improvements in congruence over time happen despite a
considerable degree of policy-making activity, as evidenced in the bottom-right panel of
Figure A1. The plot shows the share of policies in the sample that experienced policy change
41
(no matter whether favoured by public opinion or not) over four years. In the UK, around one
quarter of the policies in the sample are changed within our period of observation; in
Germany and Denmark, a little less than half. This implies that while policy change occurs
frequently, it often goes against majority public opinion, and it does not necessarily happen
for all policies that have majority support.
The varying rates of policy change across the three countries evident in the plot also
explain how the UK can be the most responsive to public opinion and the least congruent at
the same time. While the overall level of policy-making change in the UK is low, when
change happens, it is relatively responsive to public opinion support. On the contrary,
Germany changes relatively many of its policies, but the ones that lack public support are
almost as likely to be changed as ones that enjoy the support of the majority.
The analysis so far implicitly assumed that the effect of (net) public opinion is linear (on
the scale of the predictor) in the logistic regression curves presented in the top two panels. It
turns out that the effect is more complex than that. Figure A2 plots the predicted effect of
public support (left panel) and net public support (right panel) on the likelihood of policy
change as estimated by the non-linear local polynomial regression fitting (loess) function.
Figure A2. The relationship between public opinion and policy change
42
The solid black lines show the effect in the combined dataset while the dotted lines
show the effect in models estimated on country-specific data. In the general case, it appears
that as public support for policy increases from 0 to about 20% (and net support increases
from -1 to about -0.5), the probability of policy change grows. However, between 20% and
50% support (and between -0.5 and +0.1 net support), the probability of change slightly
decreases. Once support passes the 50% threshold (and +0.1 net support), the probability of
policy change starts to increase again, but then for very high values of support, it decreases
once more. The patterns differ somewhat among the three countries and given the relatively
small sample size (especially at more extreme values of public opinion), we should be careful
not to overinterpret these results. But they remain suggestive as to the non-linearity of the
effect of public opinion on policy change. Because the form of the effect does not match a
simple polynomial function (such as quadratic or log), we do not attempt to model it in the
multivariate models we present below. As a precaution to nonlinearity, we also estimate the
multivariate models with the public support variable dichotomized at the 50% mark.
We should also note the difference between the effects of (absolute) public support and
net public support. While the effect on policy change of both these variables is positive, there
is an important subtlety. When we dichotomize public support into just two categories (above
and below 50%), the observed frequency of policy change is higher when a majority of the
public supports policy change (43% vs. 33%), in line with theoretical expectations. However,
the conclusion changes when we consider net public support. If we just dichotomize net
public support into less than 0 (hence, net opposition) and above zero (hence, net support), the
observed frequencies of policy change are almost exactly the same. This implies that when an
absolute majority of public support is absent but the percentage of the public supporting the
policy is still larger than the percentage opposing it, no policy change happens (from the 23
cases where there is no majority support but still there is net public support, 21 experience no
policy change and only 2 do). In sum, it is not so much the relative share of policy supporters
versus opponents that matters for the likelihood of policy change, but the absolute share of
supporters from all citizens (including those without strong opinions on the issue).
43
3. Coalition conflict and policy change
Figure A3. Coalition conflict and policy change in Western Europe
Figure A4. The interaction effect between coalition conflict and public support for change
44
4. Additional logistic regression models of policy change
Table A2. Additional logistic regression models of policy change
Model A1 replaces the public support variable with a measure of net public support instead.
Model A2 uses an alternative measure of public support calculated as percentage from
respondents excluding 'don't knows'. Models A3 and A4 replicate Model 2 and 3,
respectively, but with measures of coalition conflict and government preferences that take into
account unofficial coalition partners of the Danish minority government parties.
Model A1 Model A2 Model A3 Model A4
Public support 0.80 (0.30) ** 1.42 (0.55) ** 1.35 (1.00) 1.44 (0.56) **
Coalition conflict -0.02 (0.10) -0.02 (0.10) 0.00 (0.09) -0.01 (0.09)
Public support*Coal. conflict / / 0.01 (0.37) /
Government support 0.16 (0.10)` 0.17 (0.10)` 0.11 (0.08) 0.11 (0.08)
Gov. support*Public support -0.09 (0.22) -0.23 (0.41) / -0.34 (0.40)
Media saliency 0.32 (0.10) *** 0.32 (0.10) *** 0.33 (0.10) *** 0.33 (0.10) **
Existing proposal 0.97 (0.29) *** 0.96 (0.29) *** 0.98 (0.29) *** 0.99 (0.29) ***
Remaining months 0.04 (0.01) *** 0.04 (0.01) *** 0.04 (0.01) *** 0.04 (0.01) ***
Denmark 0.30 (0.32) 0.30 (0.32) 0.39 (0.31) 0.37 (0.31)
United Kingdom -0.99 (0.35) *** -0.99 (0.35) *** -0.88 (0.34) ** -0.88 (0.34) **
Intercept -3.07 (0.50) *** -3.07 (0.50) *** -3.09 (0.48) *** -3.10 (0.49) ***
AIC 517.30 518.20 519.38 518.68
Notes: Logistic regression models (with logit link). Dependent variable: occurrence of policy change. Unit of
analysis is a government spel; N=525. Unstandardized and unexponentiated coefficients. Public opinion centered
at 0.5. Government positions centered at 5.0. Media saliency is logged. Significance levels: 0 < *** < 0.001 < **
< 0.01 * < 0.05 ` < 0.1
45
Model A5 adds the absolute level of government positions and an interaction with public support. Model A6 is a
multi-level model with random effets at the policy issue (case) level. Model A7 replicates Model 3 with robust
standard errors and Model 8 with standard errors clustered at the policy issue (case) level.
Model A5 Model A6 Model A7 Model A8
Public support 1.69 (0.58) *** 1.71 (0.70) * 1.48 (0.56) ** 1.48 (0.58) **
Coalition conflict -0.05 (0.10) 0.10 (0.10) -0.02 (0.09) -0.02 (0.10)
Government support 0.16 (0.09)` 0.25 (0.11) * 0.16 (0.09)` 0.16 (0.09)`
Gov. support*Public support / -0.51 (0.51) -0.32 (0.41) -0.32 (0.41)
Government positions (right) -0.17 (0.10)` / / /
Gov. positions*Public support -0.82 (0.45)` / / /
Media saliency 0.32 (0.10) *** 0.35(0.11) *** 0.33 (0.11) *** 0.33 (0.11) ***
Existing proposal 0.96 (0.29) *** 1.21 (0.41) *** 0.96 (0.29) *** 0.96 (0.31) ***
Remaining months 0.04 (0.01) *** 0.04 (0.01) *** 0.04 (0.01) *** 0.04 (0.01) ***
Denmark 0.50 (0.33) / 0.33 (0.33) 0.33 (0.34)
United Kingdom -0.90 (0.36) *** / -0.92 (0.34) ** -0.92 (0.35) **
Intercept -3.08 (0.50) *** -3.52 (0.54) *** -3.03 (0.49) *** -3.03 (0.50) ***
AIC 513.36 530.30 519.38 519.38
46
5. Interaction effect between public and government support
Figure A5. The interaction effect between public opinion and government support
The interaction effect is based on the estimates of Model 3 in the main text. The shaded
regions indicate 75% (thus, not the more usual 95%) confidence limits. On the left panel,
public opinion runs from no support to full support on the x-axis and the two regression
slopes are drawn for issues with (+3) and without government support (-3). On the right
panel, government support run from -3 to 3 on the x-axis and the two regression slopes for the
effects on policy change (represented on the y-axis) are drawn for issues that have total public
support and no support at all.
47
6. Time until policy change: bivariate analysis
Figure A6. Kaplan-Meier survival curves: share of policy issues on which policy change is
enacted over time, per different types of proposals
48
The plots trace the share of non-adopted proposals for different subgroups of cases over each
month of the observation period (up to 48 months). Note that the starting observation date for
each case – the date of the opinion poll – is to some extent arbitrary. However, there should
be no systematic bias across the major variables of interest in terms of the starting dates. The
different panels of the figure show the survival curves for cases with majority public support
or not; decided under left/GAL and right/TAN governments; with implied policy direction
consistent with the government position or not; under higher or lower than average coalition
conflict; for cases with pre-existing bill or not, and for cases of higher or lower than average
salience. The steeper the slopes of the lines, the faster the specific types of cases represented
by the lines get adopted. The greater the distance between the two curves on each plot, the
bigger the difference in the relative share of adopted cases from the two respective groups of
cases at the point of time indicated by the x-axis.
We can see from the plot in the top-left corner of Figure A6 that during the first 8
months of the lifetime of cases, majority public support does not make a difference on the
probability of policy change enactment. After this point, a difference appears that grows until
about month 30, and then starts to shrink somewhat. The dynamic is similar for cases that
differ in implied direction of the policy changed (in line with government positions or not).
Looking at the difference between governments with left/GAL and right/TAN positions
(top-right panel), the former appear to enact a higher share of cases (consistent with the
insights from the previous analyses), and the difference grows over time. The effect of
coalition conflict sets in a few months earlier, but starts to disappear after month 16. The
effect of having an existing proposal is greatest in terms of size (bottom-left panel), but it is
mostly due to changes that accumulate in the first twelve months of the lifetime of cases: if a
policy change has not been enacted within this period, it makes little difference whether an
existing bill has been in place or not in the beginning of the observation period. The (smaller
in absolute size) effect of media salience also sets in early and starts to decline after month 18
or so. Altogether, Figure A6 shows that the effects of all variables considered so far are not
very stable over time, and most effects are rather small altogether.